Design, synthesis, molecular modeling and anti-hyperglycemic evaluation of phthalimide-sulfonylurea hybrids as PPARγ and SUR agonists

Article abstract of DOI:10.1016/j.bioorg.2019.103115

New series of phthalimide-sulfonylurea hybrids were prepared and examined for their in vivo anti-hyperglycemic activities in STZ-induced hyperglycemic rats using glibenclamide as a reference drug. Compounds 6_c, 6_d, 6_g, 6_h, 6_j and 6_k induced significant reduction in the blood glucose levels of diabetic rats ranging from 24.43 to 21.43%. Moreover, molecular docking and pharmacophore approaches were carried out to examine binding modes and fit values of the prepared compounds against PPARγ and SUR, respectively. Compounds 6_c, 6_d, 6_j and 6_m exhibited the highest binding free energies against PPARγ. Compounds 6_c, 6_j, 6_k, 6_l, and 6_n showed the highest fit values against the generated pharmacophore model. Also, QSAR technique was carried out to estimate the proposed PPARγ binding affinities and insulin-secreting abilities. The synthesized compounds showed promising estimated activities. In-silico ADMET studies were performed to investigate pharmacokinetics of the synthesized compounds. They showed considerable human intestinal absorption with low BBB penetration.

Full text of DOI:10.1016/j.bioorg.2019.103115

Accepted Manuscript
Design, synthesis, molecular modeling and anti-hyperglycemic evaluation of
phthalimide-sulfonylurea hybrids as PPARγ and SUR agonists
Mohamed A. El-Zahabi, Eman R. Elbendary, Faida H. Bamanie, Mohamed F
Radwan, Salah A. Ghareib, Ibrahim H. Eissa
PII:
S0045-2068(19)30754-0
https://doi.org/10.1016/j.bioorg.2019.103115
103115
DOI:
Article Number:
Reference:
YBIOO 103115
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
11 May 2019
29 June 2019
9 July 2019
Please cite this article as: M.A. El-Zahabi, E.R. Elbendary, F.H. Bamanie, M.F. Radwan, S.A. Ghareib, I.H. Eissa,
Design, synthesis, molecular modeling and anti-hyperglycemic evaluation of phthalimide-sulfonylurea hybrids as
PPARγ and SUR agonists, Bioorganic Chemistry (2019), doi: https://doi.org/10.1016/j.bioorg.2019.103115
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design, synthesis, molecular modeling and anti-hyperglycemic evaluation of
phthalimide-sulfonylurea hybrids as PPARγ and SUR agonists
a
b
c
Mohamed A. El-Zahabi *, Eman R. Elbendary , Faida H. Bamanie Mohamed F
d
e
a
Radwan and Salah A. Ghareib , Ibrahim H. Eissa *,
a
Medicinal Chemistry department, Faculty of Pharmacy (Boys), Al-Azhar University,
Cairo, 11884, Egypt
b
Medicinal Chemistry department, Faculty of Pharmacy, Mansoura University, Mansoura
3
5516, Egypt
c
Biochemistry Department, Faculty of Medicine, King Abdulaziz University, Jeddah
1589, Saudi Arabia
2
d
Department of Pharmaceutical Chemistry, Faculty of Pharmacy, King Abdulaziz
University, Jeddah 21589, Saudi Arabia
e
Pharmacology department, Faculty of Pharmacy, King Abdulaziz University, Jeddah
2
1589, Saudi Arabia
*
Corresponding authors:
Mohamed A. El-Zahabi
Pharmaceutical Chemistry Department, Faculty of Pharmacy (Boys), Al-Azhar
University, Cairo, 11884, Egypt
Email: malzahaby@yahoo.com
Ibrahim H. Eissa
Pharmaceutical Chemistry Department, Faculty of Pharmacy (Boys), Al-Azhar
University, Cairo, 11884, Egypt

Email: Ibrahimeissa@azhar.edu.eg
Abstract
New series of phthalimide-sulfonylurea hybrids were prepared and examined for their in
vivo anti-hyperglycemic activities in STZ-induced hyperglycemic rats using
glibenclamide as a reference drug. Compounds 6 , 6 , 6 , 6 , 6 and 6 induced significant
c
d
g
h
j
k
reduction in the blood glucose levels of diabetic rats ranging from 24.43 to 21.43 %.
Moreover, molecular docking and pharmacophore approaches were carried out to
examine binding modes and fit values of the prepared compounds against PPARγ and
SUR, respectively. Compounds 6 , 6 , 6 and 6 exhibited the highest binding free
c
d
j
m
energies against PPARγ. Compounds 6 , 6 6 6 and 6 showed the highest fit values
c
j, k, l,
n
against the generated pharmacophore model. Also, QSAR technique was carried out to
estimate the proposed PPARγ binding affinities and insulin-secreting abilities. The
synthesized compounds showed promising estimated activities. In-silico ADMET studies
were performed to investigate pharmacokinetics of the synthesized compounds. They
showed considerable human intestinal absorption with low BBB penetration.
Keywords: Phthalimide; Anti-diabetic; Sulfonylurea; Docking; Pharmacophore; QSAR;
ADMET.

1
. Introduction
Type 2 diabetes affects approximately 200 million people worldwide, including
more than a quarter of elderly living in developed countries. Diet and exercise are first
line treatments along with oral hypoglycemic drugs to achieve the goal of improving
glycemic control and preventing both microvascular and macrovascular complications [1,
2
]. Sulfonylurea receptors (SUR) and peroxisome proliferator-activated gamma receptors
(PPARγ) are the main molecular drug targets (receptors) involved in the treatment of
type-2 diabetes mellitus [3, 4].
Sulfonylureas were the mainstay of anti-diabetic therapy. The second generation
sulfonylureas (glipizide I and glibenclamide II) are widely employed worldwide [5].
These drugs exert their anti-diabetic effect by stimulating insulin secretion from the
pancreatic β-cells [6-8]. The sulfonylureas bind with high affinity to the sulfonylurea
receptor (SUR), block ATP-sensitive K+ (KATP) channels in the β-cells, and stimulate
insulin secretion at both low and high glucose concentration [9, 10]. This glucose
independent insulinotropic effect of sulfonylurea often leads to hypoglycemia [11, 12].

Therefore, agents which induce insulin secretion augmented by desired glucose level will
be highly appreciated for the treatment of type 2 diabetes.
Moreover, they produce clinical effect by acting through not only SUR but also
PPARγ [13-15]. It was reported that glimepiride and glibenclamide activate endogenous
PPARγ in adipocytes. Tolbutamide was reported to exhibit PPARγ binding pattern[14] .
So that, some sulfonylurea compounds can activate both SU and PPARγ receptors.
Pharmacophore model for sulfonylurea acting drugs was designed to comprise three
lipophilic centers separated by amidic and anionic linkers [16] (Fig. 1A).
Peroxisome proliferator-activated receptors gamma (PPARγ) is a well-
characterized member of the PPAR subfamily [17]. It has a crucial role in diabetes
treatment through increasing insulin sensitizing by different tissues [18]. The essential
pharmacophoric features of PPARγ agonist are an acidic head, a linker attached to an
aromatic scaffold (spacer group), another linker, and a hetero-aromatic lipophilic tail
(Fig. 1B) [19]. Thiazolidinediones (TZDs) such as rosiglitazone III, pioglitazone IV are
well-known PPARγ agonists (Fig. 1B) [20, 21] .
In addition some phthalimide derivatives V and VI [22] have been reported to
exhibit strong and hypoglycemic effect (Fig. 1C). So, in this work, we report the
synthesis of new phthalimide derivatives bearing sulfonylurea moieties, with the aim of
reaching new compounds with dual agonistic activities against PPARγ and SUR. It worth
mentioning that Weber et al [23, 24] had synthesized phthalimide derivatives bearing
sulfonylurea moieties having hypoglycemic effect, but unfortunately they didn’t reported
the PPARγ agonistic activities of the synthesized compounds.

A
Iipophilic
Center
Amide
Linker
Anionic
linker
Iipophilic
Center
Iipophilic
Center
CH3
O
S
O
O
O
N
N
O
N
O
O
H
H
O
NH
H C
3
N
HN
N
S NH
O
H
Cl
Glipizide I
Glibenclamide II
Acidic
B
Cyclic
Tail
Aromatic
center
Linker
Linker
Head
O
O
CH3
H C
3
NH
O
NH
O
N
S
S
O
N
O
Rosiglitazone III
Pioglitazone IV
O
N
N
N
O
O
N
N
Cl
Cl
Cl
Cl
C
O
V
VI
Fig. 1: A) Pharmacophoric features of sulfonylurea agonists B) Pharmacophoric
features of PPARγ agonists C) Reported anti-hyperglycemic phthalimides
1
.1. Rationale
As an extension for our previous works [25, 26], and depending on ligand based
drug design approach [27-29], we modified the well-known drug; glibenclamide to
produce potent anti-hyperglycemic agents with dual agonistic activities against PPARγ

and SUR. The modified compounds were designed to possess the essential
pharmacophoric features of SUR and PPARγ agonists.
The acidic head required to PPARγ agonistic activity, has been replaced by
sulfonylurea moiety. They form hydrogen bonds with influencing residues of the protein
active site (Ser289, His323, His449 and Tyr473) [30]. Aromatic and aliphatic
substitutions on sulfonylurea moiety act as a lipophilic center required to SUR agonists.
Moreover, NH group of sulfamoyl moiety is acidic (pKa = 4.9 – 6.5) and there for
completely ionized at physiological pH [31]. This ionization produces an anionic linker
which contributes significantly to SUR agonistic affinity [16]. Sulfonyl (SO ) group acts
2
as one atom spacer between an acidic head and aromatic group. This is considered as
optimal length for PPARγ agonistic activity. Para-disubstituted phenyl group acts as an
aromatic scaffold (spacer group) which is required for optimal PPARγ agonism [32], in
addition to its role as lipophilic center for SUR agonists. Several linkers between
lipophilic tail and an aromatic scaffold (spacer group) have been synthesized to study
SAR of the new compounds. Finally, phthalimide nucleus was used as lipophilic tail and
lipophilic center required for PPARγ and SURs agonistic activities, respectively.
In this work, it was decided to modify its structure by such a way that the amidic
function is incorporated to a bulky heterocyclic system, namely phthalimide nucleus, with
the aim to explore the effects of this modification on the hypoglycemic action. One more
important point deserves mentioning is substituting the free rotating amide function of
glibenclamide prototypic compound with the restrictedly rotated imide nitrogen which
will be a part of the phthalimide moiety. It is of interest also to try zero, one or two
methylene groups as a spacer between the phenyl sulfonyl group and the phthalimide ring
in order to study the effect of the different conformational orientation on the
hypoglycemic activity and shed some light about the structure activity relationship of

such interesting design and promising class of compounds. Moreover, a chloro group at
the 3-position of the phthalimide moiety will be important by mimicking the
glibenclamide and would be important to compare with the unsubstituted ring analogs.
Toward that goal, it was thought worthwhile to design and synthesize three novel analogs
(Targets A, B and C) (Fig. 2) to evaluate them as hypoglycemic agents.
A typical sulfonylure
a agonist
Anionic
linker
Iipophilic
Center
Iipophilic
Center
Iipophilic
Center
Linker
O
H
O
O
N
O
N
R
X
S NH
O
6a-n
Target A
Cyclic
Tail
Aromatic
center
Acidic
Head
Linker
Linker
A typical PPAR agonist
Chemical modification
O
N
O
X
O
O
O
H
X
N
R
N
S NH
O
O
O
S
O
HN R
O
HN
Target C
Target B
Fig. 2: The synthesized targets A, B and C with basic pharmacophoric features of
SUR and PPARγ.

2
. Results and discussion
2
.1. Chemistry
The synthesis of the target compounds is depicted in Schemes 1-3. Two methods
were adopted to synthesize the starting key compounds 3 [33-35]. The first one was
a-f
reported by Santos et al [33] where, phthalic anhydride was allowed to react with
benzene sulfonamide derivatives namely, 4-aminobenzenesulfonamide 2 , 4-(amino
a
methyl)benzene sulfonamide 2 and 4-(2-aminoethyl)benzene sulfonamide 2 in acetic
b
c
acid. This method faced some problems such as long reaction time (24 h) and the low
yield of the product. A modification was adopted to prepare the desired key compounds
3a-f via substituting acetic acid by DMF and Na CO mixture, thus reduced the reaction
2
3
time from 24 h to few hours and pouring the reaction mixture on acidic ice-cold water
instead of evaporation under vacuum. Such modification showed better yields than the
reported procedure. Anhydrous Na CO was used in this method because 4-
2
3
aminomethylbenzenesulfonamide 2 and 4-(2-aminoethyl) benzene sulfonamide 2 were
b
c
available as HCl salts (Scheme 1)
Treating compound 3 with ethyl chloroformate in refluxing acetone in the
f
presence of catalytic amount of anhydrous Na CO afforded the intermediate 4 in poor
2
3
yield (40 %), which upon reaction with cyclohexyl amine in toluene resulted in the target
1
product 5 in poor yield too (30 %) (Scheme 2). The H NMR spectrum of 4 showed a
typical quartet-triplet pattern for the ethyl ester, the triplet signal appeared at 1.05-1.08
ppm while the quartet one was observed at 3.95-3.99 ppm. Also, this group disappeared
1
in the H NMR spectrum of compound 5 and the protons of cyclohexyl amine moiety
appeared in the expected regions of the spectrum. Compound 5 was synthesized by
another route reported through a one pot reaction. In this method, the sulfonamide 3 was
f
allowed to react with Na CO in acetone overnight then the reaction mixture was treated
2
3

with cyclohexyl isocyanate to get the target compound in good yields (76%)[23, 24]
(
Scheme 2).
Consequently, compounds 6 were synthesized in analogy to the one pot reaction
a-n
(Scheme 3).
Acetic acid
O
O
S NH2
(
)n
O
X
H N
2
X
O
O
N ( )n
O
O
1a,b
2a-c
DMF/ NaHCO3
3a-f
SO NH2
2
a) X = H, n = 0
a) n = 0
b) n = 1
c) n = 2
a) X = H
b) X = Cl
b) X = H n = 1
,
c) X = H, n = 2
d) X = Cl, n = 0
e) X = Cl, n = 1
f) X = Cl, n = 2
Scheme 1: Synthesis of target compounds 3a-f
O
Pathway 1
O
O
O
O
OEt
S O
NH2
Cl
Cl
S NH
O
1
2
3
) Acetone / K CO3
N
2
N
) ClCOOEt
) Reflux
O
O
4
3f
N
NH2
C
Pathway 2
Pathway 1
/ Toluene / reflux
O
Acetone / K CO / reflux
2
3
O
O
O
N
H
S NH
O
N
Cl
5
O
Scheme 2: Synthesis of target compound 5

O
O
X
X
1
) Acetone / K CO3
2
N ( )n
N
( )n
2) RNCO/ Reflux
O
O
O
SO NH2
2
S
O
O
3a-f
6a-n
HN
HN R
a) X = H, n = 0, R= cyclohexyl
b) X = H, n = 0, R= n-butyl
c) X = H, n = 1, R= cyclopentyl
d) X = H, n = 1, R= n-butyl
e) X = H n = 2, R= cyclopentyl
,
f) X = H, n = 2, R= n-butyl
g) X = Cl, n = 0, R= cyclohexyl
h) X = Cl n = 0, R= cyclopentyl
,
i) X = Cl n = 0, R= n-butyl
,
j) X = Cl, n = 1, R= cyclopentyl
k) X = Cl, n = 1, R= cyclohexyl
l) X = Cl, n = 1, R= n-butyl
m) X = Cl, n = 2, R= cyclopentyl
n) X = Cl, n = 2, R= n-butyl
Scheme 3: Synthesis of target compounds 6a-n
2
.2. Biology
2
.2.1. In vivo anti-hyperglycemic activity
The in vivo anti-hyperglycemic activity of the synthesized compounds was determined
against streptozotocin-induced hyperglycemic rats using anti-hyperglycemic assay [36,
3
7]. Glibenclamide as one of the most active anti-diabetic sulfonylurea drugs, was used
as positive control. The results were reported as % reduction of blood glucose level
Table 1 & Fig. 3). Glibenclamide significantly decreased the blood glucose level by
9.3%. Compounds 6 , 6 , 6 , 6 , 6 and 6 induced significant reduction in the blood
(
2
c
d
g
h
j
k
glucose levels of diabetic rats ranging from 24.43 - 21.43 %. These compounds can be
arranged in a descending order according to their relative potency to that of

glibenclamide as follows: 6 > 6 > 6 > 6 > 6 > 6 . All the tested compounds were less
k
j
d
c
g
h
potent than glibenclamide as anti-hyperglycemic agents.
2
.2.2. Oral Glucose Tolerance Test (OGTT)
Induction of diabetes significantly increased the total AUC recording 71130 Vs
1
7153 mg/dL.min (+314%) of the normal control rats (Table 2). Glibenclamide induced
significant reduction in the total AUC compared to the diabetic control recording 53320
Vs 71130 mg/dL.min. The other tested compounds produced promising change the AUC.
This means that these compounds have an extra pancreatic effect (PPARγ agonistic
activity). They affect the glucose utilization by the muscles or the liver and may exert
their effect through increasing of insulin sensitization.
Table 1: In vivo reduction in blood glucose levels for 7 days, in an equamolar doses
(5 mg/kg) of the synthesized compounds and glibenclamide.
Blood glucose level (mg/dL)a
Reduction
in blood
glucose
level
Reduction
in blood
glucose
Relative
potency to
glibenclamide
(RP)
Rats and
treatment
Before
After
treatment
treatment
level (%)
(mg/dL)
Normal Rats
Diabetic Rats
116.1±11.08
471.2±53.29
105.5±7.91
450.8±28.79
332.8±25.23
301.2±10.64
319.3±15.20
338.9 ±10.20
10.6
9.20
4.32
-
20.4
137.8
58.7
66.6
51.4
-
Glibenclamide 470.6±22.77
29.28
16.31
17.27
13.42
1
5
359.7±9.10
385.6±12.19
390.3±10.19
0.56
0.59
0.46
6a
6b

6c
6d
6e
6f
447.8±47.55
498.2±19.84
501.4±6.816
487.8±30.05
263.1±74.99
271.8±70.27
582.4±12.54
551.4±6.816
525.6±20.616
376.1±18.71
466.75±7.66
503.1±40.80
352.2±63.03
390.6±22.58
402.4±20.10
447.2±45.58
207.2±11.65
214.1±84.08
521.6±18.75
422.4±20.10
397.2±13.52
301.3±15.60
416.7±10.25
450.8±26.86
95.6
107.6
99.0
40.6
55.9
57.7
60.8
129.0
128.4
74.8
50
21.43
21.59
19.74
8.32
0.73
0.74
0.67
0.28
0.73
0.73
0.36
0.80
0.83
0.67
0.37
0.35
6g
6h
6i
21.24
21.22
10.43
23.39
24.43
19.88
10.71
10.39
6j
6k
6l
6m
6n
52.3
a
All data within 10 % (n = 3)
3
3
2
2
1
1
5
0
5
0
5
0
5
0
Rats and
treatment
Fig. 3: Antihyperglycemic effect of the tested compounds

Table 2: Total area under the curve (AUC) after glucose loading in normal and
diabetic rats pretreated with vehicle, glibenclamide and the synthesized compounds
in an equimolar dos of 5 mg/kg.
Total AUC (mg/dl.
min)
Reduction in AUC
(%)
Treatment
1
7
7153 ± 123.33
1130 ± 326.24
NRa
NRa
Normal rats
Diabetic Rats
Glibenclamid
e
53320 ± 432.23
2
5
5
7400 ± 5423.27
6440 ± 6534.25
6
19
20
17
17
22
19
18
c
5
6d
6g
6h
6j
5
8810 ± 2354.12
5
8610 ± 5423.17
5260 ± 3329.17
5
6k
6l
57350 ± 4012.32
8240 ± 4135.38
5
a
No reduction in AUC
2
.3. Molecular modeling
Depending on availability of PPARγ crystal structures in the Protein Data Bank,
receptor based drug design (docking) technique [38-42] was used to investigate the
binding pattern of the synthesized compounds with the prospective target, PPARγ. On the

other hand, the crystal structure of SUR is not available, so that, ligand based drug
design (pharmacophore) technique [43, 44] was used to predict proposed agonistic
activity of the synthesized compounds against SUR.
2
.3.1. Docking studies
Crystallographic structure of PPARγ was retrieved from Protein Data Bank [PDB ID-
FM6, resolution 2.1 Å] (http://www.pdb.org), and considered as a target for docking
1
simulations. The docking analysis was performed using MOE [45] software to evaluate
the free energies and binding mode of the designed molecules against PPARγ. At first,
the crystal structure of PPARγ was prepared by removing water molecules and retaining
only one chain and its co-crystallized ligand, rosiglitasone. Then, the protein structure
was protonated, and the hydrogen atoms were hided. Next, the energy was minimized,
and the binding pocket of the protein was defined.
The 2D structures of the synthesized compounds and reference ligands
(rosiglitazone and glibenclamide) were sketched using ChemBioDraw Ultra 14.0 and
saved as MDL-SD format. Then, the saved file was opened using MOE and 3D structures
were protonated. Next, energy minimization was applied. Before docking the synthesized
compounds, validation of the docking protocol was carried out by running the simulation
only using the co-crystallized ligand and low RMSD between docked and crystal
conformations. The molecular docking of the synthesized compounds and the co-
crystallized ligand was performed using a default protocol. In each case, 10 docked
structures were generated using genetic algorithm searches
The PPARγ cavity consists of three main parts; an entrance, arm I and arm II. The
arm I contains four polar residues, Tyr473, Ser289, His449 and His323, involved in
hydrogen bonding. Arm II comprises Leu353, Ile 341, Ile281 and Val339 while the
entrance consists of Arg288, Ser342, Leu330and Leu333 [30] (Fig. 4).

Fig. 4: PPARγ cavity composed of three main parts arm I, arm II and entrance [30]
The results of docking studies revealed that the synthesized compounds exhibited
similar orientations inside the putative binding sites of PPARγ. The designed members
showed good binding energies ranging from -14.00 to -26.63 kcal/mol (Table 3).
The proposed binding mode of the co-crystallized ligand, rosiglitazone, showed
an affinity value of - 27.89 kcal/mol. Thiazolidinedione moiety was oriented in the polar
arm I of the target receptor. Hydrogen bonding interaction with a distance of 2.51 A˚ was
observed between the acidic hydrogen of the imide group and Ser289. The nitrogen atom
of pyridine group formed one hydrogen bond with Ser342 with a distance of 2.34 A˚. The
pyridine moiety occupied the area of the hydrophobic pocket forming pi-cation
interaction with Arg288, besides two pi-pi and pi-sigma interactions. These results were
in agreement with the reported data [30] (Fig. 5).

Compound 6 as representative example exhibited a binding mode similar to that
j
of rosiglitazone, with affinity value of - 26.63 kcal/mol. The sulfonylurea moiety was
oriented in the polar arm I of PPARγ and formed two hydrogen bonds with Ser289 (OH)
and Cys285 (C=O) via C=O and NH groups of urea moiety, with distance of 2.65 and
2
.58 A˚, respectively. Phthalimide moiety was oriented in the hydrophobic entrance
forming pi-pi and pi-cation interactions (Fig. 6). Mapping surface technique was carried
out to show compound 6 occupying the active pocket of PPARγ (Fig. 7).
j
The proposed binding mode of compound 6 (affinity value of -25.61 kcal/mol)
m
was virtually the same as that of rosiglitazone, where its sulfonylurea moiety directed into
the polar arm I. It formed two hydrogen bonds with Tyr473 (OH) and Tyr327 (OH) via
NH and C=O groups of urea moiety, with distance of 2.42 and 2.77 A˚, respectively. The
heterocyclic phthalimide moiety was oriented in the hydrophobic entrance forming pi-pi
and pi-cation interactions with Arg288 (Fig. 8).
Table 3: The docking binding free energies of the synthesized compounds
Binding free
energy
Binding free
energy
Comp. No.
Comp. No
(
kcal/mol)
(kcal/mol)
- 21.75
5
- 14.19
- 14.00
- 20.41
- 25.11
- 25.12
- 21.26
- 19.63
- 14.44
- 19.54
6i
6j
6a
6b
6c
6d
6e
6f
6g
6h
- 26.63
6k
6l
- 22.64
- 23.76
6m
6n
- 25.61
- 19.84
Glibenclamide -24.11
Rosiglitazone - 27.89

Fig. 5: Binding mode of rosiglitazone forming hydrogen bond (green dot) between
its polar head and Ser289

Fig. 6: Compound 6 docked into the active site of PPARγ. The sulfonylurea moiety
j
formed two hydrogen bonds (green) with Ser289 and Cys285 residues.

Fig. 7: Mapping surface showing compound 6j occupying the active pocket of
PPARγ

Fig. 8: Compound 6 docked into the active site of PPARγ. The sulfonylurea moiety
m
formed two hydrogen bonds (green) with Tyr473 and Tyr327residues.
2
2
.3.2. Pharmacophore studies
.3.2.1. Generation of 3D-pharmacophore model
A 3D pharmacophore model was generated and validated using a set of thirty one
compounds containing a sulfonylurea scaffold [25, 26, 46] (Fig. 15). The generated
model was used as 3D look inquiry for determination the agonistic activities of the
synthesized members against SUR.
In general, ten hypotheses were generated as shown in Table 4. Most of hypotheses
involved the following features; Hydrogen bond acceptors (HBA), hydrogen bond donor
(HBD), hydrophobic aliphatic (HA) and ring aromatic (RA). All of the generated
hypotheses have cost differences over than 60 bits and high correlations indicating their
stability. The total cost of some hypotheses is close to the fixed cost and very different to
the null cost value, indicating good hypotheses.

Hypo-1 (the one with the highest correlation, the lowest RMS and the lowest total
cost) was the optimal one and used for further analyses. The RMS value of 1.01,
correlation coefficient of 0.89 and cost difference of 115.14 (Table 4). It consisted of
four features (HBA, HBD, HA, RA). The spatial arrangement and the distances between
the pharmacophore features of hypo-1 are demonstrated in Fig. 9.
Table 4: . Statistical significance, predictive power and features of the top-10
pharmacophore hypotheses derived from the training set of molecule.
Total
cost
Cost
Max.
fit
Hypo.
RMSb Correlation Featuresc
differencea
Hypo 1
Hypo 2
Hypo 3
Hypo 4
Hypo 5
Hypo 6
Hypo 7
115.14 96.70
117.77 94.07
119.14 92.70
122.37 89.47
124.69 87.15
121.73 90.11
122.61 89.23
1.01
1.02
1.27
1.63
1.10
1.14
1.21
0.89
0.78
0.62
0.61
0.59
0.72
0.67
HBA, HBD, HA, RA
HBA, HBD, HA, RA
HBA, HBD, RA
9.35
8.22
6.53
HBA, HBA, HBA, HAr, RA 5.01
HBD, HBD, HBD, HA
HBA, HBA, HA, RA
HBA, HBA, HBD, HA
6.20
8.19
6.76

Hypo 8
Hypo 9
123.38 88.46
123.42 88.42
1.28
1.17
1.21
0.62
0.70
0.67
HBA, HBA, HBD, HA
HBA, HBA, HA, RA
HBA, HBA, HBD, HA
7.52
8.04
7.52
Hypo 10 123.51 96.70
a
Cost difference is the difference between null cost and total cost. Null cost = 214.84 bit.
RMS, root mean square deviation, the deviation of log estimated activities from log
b
experimental activities normalized by log uncertainty.
c
Abbreviation used for features; HBA: hydrogen bond acceptor, HBD: hydrogen bond
donor, HA: hydrophobic aliphatic, HAr: hydrophobic aromatic, RA: ring aromatic.
Fig. 9: Hypo 1 with four features; green (HBA), blue (HA), purple (HBD) and
brown (RA). The distance constraint between pharmacophore features is reported
in angstroms.
In order to assess the Hypo-1 predictability, experimental and estimated activities of
the training set compounds were compared. The results indicated that there is marked
consistency between experimental and predicted activity values for all test compounds.

This indicates that the pharmacophore Hypo-1 classifies the training set compounds
correctly (Table 5).
Table 5: Experimental and estimated activities of the training set compounds based
on pharmacophore model hypo 1.
Exp.a
µM)
Estb
(µM)
0.01
0.02
0.02
0.03
0.03
0.03
0.04
0.04
0.04
0.04
0.04
0.04
0.05
Fit
Comp.
Errord
(
valuec
8.81
6.57
8.47
8.34
6.27
5.27
6.22
7.22
7.20
9.16
6.14
7.14
7.12
7a
7b
7c
7d
7e
8a
8b
9a
9b
0.01
0.01
0.02
0.03
0.01
0.08
0.03
0.03
0.03
0.04
0.02
0.03
0.04
-1.06
1.48
-1.18
-1.02
2.51
-2.30
1.10
1.26
1.33
-1.02
2.00
1.34
1.05
1
1
1
1
0a
0b
1a
1b

1
1
1
1
1
1
1
1c
1d
2a
2b
2c
2d
2e
0.03
0.03
0.10
0.09
0.09
0.10
0.04
0.05
0.05
0.05
0.05
0.05
0.06
0.06
6.10
6.10
6.10
6.07
6.06
6.03
7.01
1.45
1.92
-1.95
-1.69
-1.77
-1.73
1.36
a
b
c
Experimental activity value
Estimated activity value.
Fit value calculated by geometry fitting between the hypothesis and the compound; the
higher value, the better fit.
d
Error factor calculated as the ratio of the measured activity to the estimated activity or
the inverse if the estimated activity is greater than the measured activity. If measured
activity is greater than the estimated activity, the error factor is negative and vice versa.
2
.3.2.2. Validation of Hypo-1 pharmacophore model
Validation processes were carried out to test the ability of the built model to distinguish
active structures and anticipate their activities precisely. Two methods were done to
validate the hypo-1; test set activity prediction and mapping of glibenclamide as an
external reference of the generated pharmacophore.
2
.3.2.2.1. Test set activity prediction
Test set of ten compounds was subjected to the same protocol of training set to verify the ability
of the generated pharmacophore model (hypo-1) to predict the activities of the new synthesized
compounds. The obtained results showed that hypo-1 can classify the active and inactive compounds
correctly. Experimental and estimated activities and fit values of the test set were listed in Table 6.
Table 6: Experimental and estimated activities and fit values of the test set based on

pharmacophore model hypo-1.
Comp. Exp. (µM) Est (µM)
a
b
Fit valuec
7.99
7.98
7.97
6.97
7.97
6.96
7.95
7.93
8.87
7.82
7.76
7f
7g
8c
9c
NAd
0.09
0.09
NAd
NAd
NAd
0.03
0.04
NAd
0.09
0.10
2.56
0.06
0.07
3.17
2.22
1.97
0.07
0.07
1.08
0.09
0.10
1
1
1
1
1
1
1
0c
1e
1f
2f
2g
2h
2i
a
Experimental activity values.
b
Estimated activity values.
c
Fit value indicates how well the features in the pharmacophore overlap the chemical features in
the molecule.
d
NA: not active
2
.3.2.2.2. Mapping of glibenclamide
Glibenclamide was mapped with the generated pharmacophore (hypo-1) and demonstrated a high
fit value of 7.98. The chloro phenyl moiety completely fitted with RA, the sulfonyl group fitted with
HBA, the central phenyl ring fitted with RA and the NH of amide linker fitted with HBD (fig. 10).

Fig. 10: Mapping of glibenclamide on the generated high ranked pharmacophore
Hypo1); green (HBA), blue (HA), purple (HBD) and brown (RA).
(
2
.3.2.3. Estimation of designed molecules
The validated pharmacophore model (Hypo-1) was used to estimate the activities of the
synthesized members according to the degree of fitting with Hypo-1. The estimated activities and fit
values of the synthesized compounds were listed in Table 7. It was observed that most of the
synthesized compounds have high fit values toward Hypo1, indicating promising SUR agonistic
activities. Compounds 6 , 6 6 6 and 6 showed the highest fit values against the generated
c
j, k, l,
n
pharmacophore model Mapping of three designed compounds (6 , 6 and 6 ) on the generated
j
k
l
pharmacophore (Hypo1) was demonstrated in Fig. 11, 12 and 13, respectively. It was noted that the
mapping pattern of these compounds is as the same as that of glibenclamide.
Table 7: Expected activities and fit values of designed compounds based on the
Hypo 1 pharmacophore model
Est. (µM) Fit valueb Comp.
a
Est. (µM) Fit value
a
b
Comp.
5
0.06
0.10
0.08
7.19
5.80
7.43
6h
6i
0.03
0.08
0.03
7.60
7.32
8.79
6a
6b
6j

6c
6d
6e
6f
0.07
0.08
0.06
0.12
0.10
7.66
6.52
7.64
7.52
6.62
6k
6l
0.03
0.03
0.06
0.04
7.79
7.75
6.83
8.14
8.94
6m
6n
6g
Glibenclamide 0.02
a
Estimated activity value.
b
Fit value indicates how well the features in the pharmacophore overlap the chemical features in
the molecule.
Fig. 11: Mapping of compound 6j (fit value = 8.79) on the generated high ranked
pharmacophore (Hypo1); green (HBA), blue (HA), purple (HBD) and brown (RA).

Fig. 12: Mapping of compound 6k (fit value = 7.79) on the generated high ranked
pharmacophore (Hypo1); green (HBA), blue (HA), purple (HBD) and brown (RA).
Fig. 13: Mapping of compound 6l (fit value = 7.75) on the generated high ranked
pharmacophore (Hypo1); green (HBA), blue (HA), purple (HBD) and brown (RA).
In order to probe the similarity between the 3D structures of one of the synthesized
compounds 6 and glibenclamide; flexible alignment was carried out via molecular overlay protocol
e

using discovery studio 2.5 software [44]. Compound 6 showed a distinct similar pattern with
e
significant fairly alignment with glibenclamide (Fig. 14).
Fig. 14: Flexible alignment of compound 6 (faint blue) and glibenclamide (green).
e